Analytical procedures are only as reliable as the materials used to calibrate the procedures. These materials may be a sample without the analyte of interest (a "base-line calibration sample") or with a known amount of the analyte.
Good laboratory practice dictates the use of calibration samples prepared in a matrix identical to the matrix of the sample to be tested. With few exceptions, test accuracy will be compromised when this approach is not followed because new variables have been introduced to the test. Test methods that use extraction and immunoassay techniques are particularly sensitive to matrix composition and require proper validation of test accuracy with a designated matrix.
The usual way of preparing a calibration blank for an assay involving an exogenous analyte (such as a drug) is to supplement drug-free serum or urine with known quantities of the drug. Unfortunately, this method is problematic when the analyte is a normal endogenous biological constituent. Cortisol, for example, is a steroid present in all human sera. Adding pure cortisol to native human sera will produce an inaccurate calibration blank. Because endogenous steroids are normal constituents of serum, they must be removed from a calibration serum sample before standard amounts of a steroid can be added.
To circumvent this problem, investigators and commercial manufacturers often prepare cortisol calibration blanks in dilute albumin solutions or in a serum matrix that has been significantly altered during the process of removing endogenous cortisol. For example, McBride et al., Clinical Chem., 37[5], 1991, describes cortisol calibration blanks prepared in a dilute albumin solution, rather than in serum. Accordingly, when immunoassays are used to test for cortisol (or other steroids) accuracy problems can surface that are due, in part, to dissimilarity between the matrices of the calibration blank and the test specimen. Deviations from the usual sample environment will affect chemical interactions, analytical response, and overall accuracy. Discrepancies are most obvious when popular commercial immunoassays are evaluated against reference chromatographic/mass spectrometry methods.
Some investigators do not remove steroids but instead use a standard addition technique. With this technique a known quantity of a steroid standard is added to test sample. Both native test sample with and without added standard are analyzed. The difference between the two results represents the analytical response for the standard. The remaining response is from the steroid in the native test sample. This indirect method of standardizing is rarely used because of potential accuracy problems. (Frey, et al., ClinincaI Chem. 25[11], 1979.)
A common method of removing steroids from serum is with charcoal. Soldin et al., Therapeutic Drug Monitoring, 14:164-168, 1992, describes charcoal adsorption of steroids. This treatment is a harsh process that will nonspecifically adsorb both steroids and other endogenous compounds. It is difficult to remove residual charcoal from serum, and even trace amounts can interfere in subsequent assays that use the steroid-stripped serum as a calibration base sample. Processes used to remove charcoal can introduce contaminants into serum.
What is needed in the art of analyte analysis is a serum sample useful as a calibration base-line comprising a natural blood serum matrix.